Mehdi Hosseinzadeh
Australian Centre for Robotic Vision, Washington State University, University of Adelaide
Papers
8
Total Citations
98
H-Index
5
About
Mehdi Hosseinzadeh is a robotics and control systems researcher whose work spans simultaneous localization and mapping (SLAM), safety-critical control, and human-robot interaction. His most influential contribution, "Structure Aware SLAM Using Quadrics and Planes" (2019, 63 citations), advanced the field of semantic mapping by integrating geometric primitives into sparse SLAM frameworks, enabling robots to build richer, more structured representations of their environments. Complementing this, his work on monocular object-model aware SLAM further bridges the gap between accurate camera localization and semantic scene understanding. Beyond perception, Hosseinzadeh has made notable strides in safe autonomy. His research on explicit reference governors offers optimization-free solutions for controlling safety-critical systems under complex, time-varying constraints—a practically significant contribution for real-world deployment. His recent work on human-robot interaction introduces the compelling concept of "danger awareness," quantifying whether humans recognize and engage with nearby robots to inform safer, more adaptive robot behavior. This thread extends into action planning frameworks designed to handle careless or unconcerned humans probabilistically. With a growing body of work spanning robotic manipulation, constrained control, and socially aware autonomy, Hosseinzadeh represents an emerging voice at the intersection of intelligent robotics and formal safety guarantees.
Research Focus
Key Achievements
Top Papers
- 1Structure Aware SLAM Using Quadrics and Planes63 citations · 2019
- 2
- 3Real-Time Monocular Object-Model Aware Sparse SLAM9 citations · 2019
- 4
- 5Structure Aware SLAM using Quadrics and Planes5 citations · 2018
- 6Robot Action Planning in the Presence of Careless Humans3 citations · 2024
- 7
- 8